Development of a test method for measuring galling resistance

(1)

Karlstads universitet 651 88 Karlstad

Fakulteten för Materialteknik

Fredrik W. Lindvall

Development of a test method for

measuring galling resistance

Master thesis

Date/Term: 2007-06-06

Supervisor: Pavel Krakhmalev

(2)

Abstract

Today sheet metal forming is used to make a variety of mass production because it has a high production rate. One of the biggest concerns in sheet metal forming is wear of the tool in form of galling. Galling in sheet metal forming is characterised by an increased tool surface roughness, unstable friction in the forming process and undesirable scratches on the final products.

Several ways of ranking materials resistance to galling exist today but only ASM G98 is standardised. Nevertheless, some different methods developed for ranking tool materials’ tendency to galling have also been developed.

The aim of this thesis is to develop and improve the Uddeholm Tooling Tribo Test rig located at Uddeholm Tooling AB. The rig, which is a variation of cylinder-on-cylinder test equipment, was improved with a new tool holder, a utilization of the real sheet material counter face and a new data acquisition system and software. The galling was detected using scratches on the sheet, metallographic analysis of the material adhered on the tool specimen, monitoring of coefficient of friction and the standard deviation of the coefficient of friction.

(3)

Abstract ...2

Introduction... 4

Classification of main wear modes ... 4

Influence of microstructure and surface characteristics on wear modes ... 7

Lubrication regimes... 13

Wear and lubricants in sheet metal forming. ... 16

Galling as a particular problem ... 19

Galling mechanisms ... 19

Galling initiation and growth in deep drawing... 20

The contact important in SMF... 21

Methods for galling initiation tests ... 24

Aims and scope... 30

Materials and Methods... 31

Sample preparation... 32

The UTTT rig... 33

A model test... 33

To chose a criterion for galling ... 35

Hardware ... 37

The sheet material ... 37

The tool holder platform ... 38

The Data Acquisitioning System... 39

(4)

Introduction

Today galling in sheet metal forming is a problem that requires a closer look. Because governments enforce harder restrictions on pollution the industry is forced to change lubricants. No new lubricant has yet been able to provide the necessary wear resistance and meet the demands on being a non pollutant. Also a desire to operate with less or no

lubrication is driving research as to how materials can be engineered to have a natural resistance against galling. In this thesis a method for testing materials resistance to galling is evaluated as to its capability to rank tool steels.

Classification of main wear modes

It is commonly known that sometimes a piece of equipment needs to be replaced due to wear. It applies to almost everything, from heavy industrial machinery to everyday objects like car tiers and kitchen knives. In this section is an overview given of wear modes and general contact situations associated with wear.

Wear can be defined as loss of material from one or both surfaces due to the relative motion of two surfaces in mechanical contact with each other [1]. This is found in many different applications, for instance when the edge of a sheet cutter is dulled over time or the cam strap on a car which is needed to be replaced periodically. The cogs on cogwheels and the balls in ball bearings are also worn, in fact almost all moving parts is subjected to some kind of wear.

(5)

Fig. 1. Micromechanical mechanisms, from left to right; shear, fracture, cutting, extrusion and diffusion, [1].

and in the fracture mechanism the material is or has become brittle and is removed as small bits are broken off. In cutting an abrasive element cuts a chip just like a micro scale tool edge. In extrusion material is pushed in front of an asperity. Diffusion is mostly associated with gases and liquids but take place in solids as well. In diffusion a material in high concentration seek to even out the difference and become more uniformly dispersed, see Fig. 1, [1].

(6)

particles impact a surface and like in concrete blasting the particles can be carried by air or liquid. Erosive wear is a concern in fans operating in particle rich environments. It is also a distress in oil drilling equipment where particles are transported in liquid.

The fact that more than one micromechanical mechanism can act simultaneously

complicates the matter significantly. Here mechanisms that do not remove any material in themselves can have a large impact on the wear rate. An example of when the mechanisms work together to reduce the total wear is that the oxide layers of sliding metal bodies grows thick and separates the metal surfaces. Oxidation is able to work to counter severe

extrusion and adhesion in sliding contacts. By forming a hard layer separating the metal surfaces, the friction decreases and the wear is reduced. But more severely the mechanisms could work together and cause greatly accelerated wear, as when oxidation and abrasion act simultaneously under high temperatures. The abrasion removes the oxide layer which grows back quickly and is subsequently removed again by wear. The wear rate is many times higher than if oxidation or wear acted individually [1].

(7)

Fig. 2. Tribo-system, principal image, [2].

Influence of microstructure and surface characteristics on wear modes

In order to better understand the special characteristics of the surface a brief look at the bulk material is appropriate. The bulk material of metals has ideally an ordered crystal structure throughout the entire body. In reality the bulk material consists of areas with an ordered crystal structure stacked close together. see Fig. 3a,b. The areas of ordered structure are referred to as grains and the boundaries between them as grain boundaries. The sizes of the grains are usually determined with the help of an American Society for Testing and Materials (ASTM) issued chart. The comparison chart has images of grains at 100 times magnification and a number associated with each image. The numbers is termed

grain size number and ranges from 1 to 10 where 10 represent small grains. This number

is determined by where N is the average number of grains at 100 times magnification per square inch (1 in = 25.4 mm) and n is the grain size number [

1

2 −

= n

N

(8)

Fig. 3. a) Principal image of grains and grain boundaries. b) “Photomicrograph of the surface of a polished and etched polycrystalline specimen of an iron chromium alloy in which the grain boundaries appear dark.” [3]. c) Atoms and their bonds in both the bulk and the surface. The black

dots represent atoms and the lines bonds. In the bulk all atoms use all their bonds while in the surface some bonds are free. This makes the surface especially reactive. d) The layered structure of

a technical surface. Fig. 3a and 3b are taken from [3]. 3c and 3d are taken from [1]. 3d have been adapted by the author

These free bonds make the surface particularly reactive. Carbon alkanets and water are especially prone to be adsorbed and quickly form cohesive surface layers even at low concentrations [1].

In effect the majority of surfaces are layered, see Fig. 3d. Almost all manufacturing require some mechanical processing to achieve a desired form and surface. Furthermore, surfaces characteristics produced by machining and wear are similar. During these

(9)

and chemical composition are created. The grains closest to the surface are broken down to an extremely fine size (< 0.1 µm) [1]. These mechanically processed layers are almost always harder than the bulk material. The depth of the deformed zone depends on how rough the machining has been. In a few instances such as rough polishing material up to 1 mm deep can be affected. Some surface treatments leave no mechanical deformation, for example chemical etching [1].

On top of the mechanically deformed layer a layer of oxide is formed and its thickness depends on the type of material, machining method, atmosphere and temperature. The oxide often helps to prevent severe wear but can under unfortunate conditions accelerate it. The next and final layer consists of more or less random molecules that have reacted with the surface. Surfaces are always reactive which mean that contaminating atoms or

molecules easily adhere to the surface. This tendency of surfaces to adhere molecules can be used for engineering of new lubricants so that, desired molecules are adhered preferably over undesired [1].

Since the surface layers have a decisive effect on how the material interacts with its environment they are manipulated by finishing or altered by coatings to suit their

applications better. Because adhesion between layers on opposing surfaces is much lower than bare metal to metal contacts an effort is usually made to avoid the later [1].

Superficially a flat surface is just flat. But a closer look at this flat surface will reveal a landscape of jagged mountains and deep valleys. In fact, a close enough look and even a polished surface has valleys and asperities. This roughness on a microscale consists of two components, a slow moving waviness that is often periodically repeated and a

(10)

Fig. 4. a) Surface profile; from top to bottom, surface profile, surface wariness, surface roughness. The surface profile consists of the surface waviness with the surface roughness superimposed.

b) Surface profiles with the same Ra value, [1].

surface profile, see Fig. 4a. Today the most sensitive instruments can measure the topography of individual atoms. Atoms have a diameter of 2-5 Å (1 Å = 10-10_{m). In} practice are instruments that have a resolution of around 10 Å more common [1]. There are several parameters describing surface topography, among them are Ra, Rmax and Rp. A measurement widely used is the Ra value, the mean vertical deviation from the mean level of the surface, see Fig. 5. Even though Ra is calculated as an integral it is usually approximated with a sum when derived from digital data. Ra is a rather rough value but is nonetheless often presented along with other surface parameters. The reason Ra is such a rough parameter is that it takes no heed to the shape and form of the valleys and asperities. This is clearly illustrated in Fig. 4b, where several surfaces that have the same Ra are presented.

(11)

(12)

The coefficient of friction is assumed to be independent of the surface topography but this is a truth with modifications. If two surfaces are relatively close in surface roughness this can be taken as true, however if the surfaces differ a lot from each other then the coefficient of friction depend on the surface topography. If a smooth metal block is slid down a smooth metal surface, a coefficient of friction is measured. However if a metal block with a grinding paper glued on is slid down a smooth metal surface then the coefficient of friction will be quite another [1].

The majority of metals form an oxide layer on the surface. These oxide layers are chiefly harder than the bulk material and thickness of the layer depends on the metal chemistry, environment and temperature. Oxides in general tend to reduce wear by separating the metals from each other and by increasing the hardness of the surface. The relation between the hardness of the oxide and the bulk material has large effect on wear. If the oxide is much harder than the bulk material the bulk material will plasticize before the oxide, which results in cracking and damage of the oxide layer. The damaged oxide is then removed mechanically by the contacting bodies and new oxide grows back. The growing, damaging, removing cycle greatly accelerates wear. A thick layer of oxide can cause tensions because the oxide has higher or lower molecular volume than the bulk material. These tensions can crack the oxide and make it more prone to further damage [1].

(13)

but on the asperity level there are deformations and indentations and hardness is therefore important.

This is the main reason why most tool steels have high hardness. Hardness is usually measured by the indentation of an indenter that is pressed against the surface. The area of the indentation is measured and divided with the pressure applied to cause the indentation, thus giving hardness the dimension of [N/m2]. There are several ways to do these

measurements and among the most popular are the Vickers, Brinell, Meyer and Knoop methods [1].

What has been mention above can be said to be part of the surface quality. A high surface quality results in low wear. Smooth surfaces with a good surface profile, a stable oxide and a high hardness are often associated with low friction and low wear. High wear is almost always accompanied by a high coefficient of friction as a relatively large amount of material is sheared, cut, extruded, etc. during the process. Therefore it is important that the coefficient of friction remains low.

Lubrication regimes

(14)

Fig. 6. Lubrication regimes, a) boundary lubrication b) mixed lubrication c) hydrodynamic lubrication, [1], the figure has been adapted by the author.

the geometry of the bodies can be designed to help the lubricant flow in between [1]. Boundary lubrication does not separate the surfaces but creates a transfer film. This film is built by polarised molecules that have a “head” that binds strongly to metal surfaces and a “tail” that slides easily against it self, see Fig. 7. These molecules form a coherent carpet on both surfaces with heads to the surfaces and tails out. The surfaces slide on these low friction carpets. The lubricant is there only as a mean to get these molecules to the surfaces and to make sure that there are molecules that can replace damaged or missing ones. Of the lubricant the active molecules is only a few percent [1].

Sometimes some of the pressure is carried by the lubricant but the surfaces are still in contact with each other. Here added molecules also form surface films that promote low friction for the parts of the surface that are in contact with each other. This is called mixed lubrication [1].

Wear is reduced in hydrodynamic lubrication as the surfaces are separated and in

boundary lubrication by the low friction transfer film. Lubricants also removes wear debris from the contact zone and prevents particles to enter. Other benefits are that the

temperature is kept down and corrosion is reduced [1].

(15)

Fig. 7. Boundary lubrication and transfer films, [1].

be painted or tapped. A good lubricant adheres strongly to the surface, especially if a boundary lubricant is used, and is hard to clean. To get rid of the lubricants the products are washed with strong cleaners, and these are often environmentally unfriendly. In demanding applications even the lubricant might be hazardous [5].

In an effort to use more environmentally friendly lubricants some studies have been done on coconut oil and boric acid. In India, the coconut oil is used as engine oil in 2-stroke engines. The coconut oil has poorer wear resistance compared to commercially available oils but is in abundant supply in the area. Additives and modifications to the oil might render it an environmentally friendly alternative to mineral oils [6].

Boric acid or orthoboric acid H3BO3 is a solid laminar lubricant. Much like graphite is boric acid made of layers that easily slide over each other. The boric acid is

environmentally friendly and is an accepted engineering material mainly used in the production of glass. It can also be found in mild antiseptics and eyewash. When added to transmission fluid it reduced friction and wear. More important though is that it can be mixed with canola oil extracted from rape seeds to form an environmentally friendly lubricant. The combination with canola oil and boric acid has been found efficient in sheet metal forming [7,8].

(16)

removed and replaced with a forming lubricant. The forming lubricant must then be removed in order provide better adhesion of the primer. Expensive and cumbersome machinery are required to both apply and remove the forming lubricant. If a lubricant can protect the sheets from corrosion during storage and lubricate during the deep drawing and allow the primer to adhere properly without cleaning, it would save a lot of time and money. These are referred to as PLP-lubricants (Preservation, Lubrication, and Primer-compatibility) and have been found effective in studies made by [9].

Wear and lubricants in sheet metal forming.

Sheet metal forming (SMF) is an effective way to produce large quantities of products. It is a metal forming operation in which a sheet is plastically deformed in one or more steps. Since this can be done with various techniques the term SMF is wide and has several sub classes, including spinning, stretching, bending, deep drawing, etc, see Fig. 8.

SMF is well suited for mass production even though the initial costs are high as specific machinery is needed and the punch and die are expensive to manufacture. The strong points of SMF are that production rate is high and the cost per product is low.

In sheet metal forming the die is an “image” of the product being formed. This makes the die product unique and when two different shapes are to be produced then two different dies must also be manufactured. The die is expected to produce up to a million products and must be flawless. Any imperfection will be present in every product. The high

demands on perfection make fabrication of the die expensive. This also means that if there is any change in the die during the production, this change will be represented in the subsequent products. Thus, any significant wear of the die must be avoided.

(17)

Fig. 8. Sheet metal forming, principal sketches of different techniques, [10].

(18)

Fig. 9. Deep drawing, principal sketch, 1) punch or die, 2) clamp, 3) bank, 4) sheet, [15].

relative motion between the surfaces in relation to the load. The transfer films perform poorly under the high pressures and when the sheet is extensively plastically deformed. A way round this problem is to use stronger and more aggressive lubricants. Some common lubricants in demanding SMF operations are chloride, sulphur and zinc based lubricants. These lubricants adhere strongly to the surfaces and are quick to replace missing molecules but they are hazardous to humans and the environment. Since lubricants designed for high adhesion are hard to remove, the removal of the lubricant after SMF is difficult and requires hazardous solvents. The necessary cleaning of the products is expensive because all the residues need to be removed. This is not easy as the lubricating film adheres strongly to the sheet metal surface. Strong cleaners are used to remove the hazardous lubricant and the waste needs to be taken care of. The lubricants, cleaners and waste handling are expensive. Additionally, stricter regulation of environmentally unfriendly products has limited the use of these products today. The costs have forced the industry to look for alternatives [5].

(19)

seizure. In this case the tool needs to be re polished in order to remove the adhered material [5].

Galling as a particular problem

Galling is a somewhat loose definition and is often described differently depending on the author. Here the chosen definition of galling is “a form of surface damage arising between sliding solids, distinguished by macroscopic, usually localized, roughening and creation of protrusions above the original surface” [11] (quotation found in [17]). The phenomenon is sometimes also described as seizure, seizing, scoring or scuffing.

Galling mechanisms

As galling is not entirely understood, scientists offer more than one explanation to this phenomenon. Here follows a short review of the most popular explanations.

Since galling is more likely to occur when the materials are similar in chemical

composition a theory based on mutual solubility was suggested. This theory basically says that the higher mutual solubility the two materials have, the more likely they form strong adhesive bonds. This makes them more prone to galling [12]. Another but closely related theory correlates the crystal structure to the forming of strong adhesive bonds. In this approach the same materials but with different crystal structure have been tested [12]. One more theory assumes that metal in direct contact with metal will form strong adhesive bonds. This theory only offers a part of the explanation. Direct contact term means that there are no oxide layers or any other contaminant between the surfaces and the distance between the oxide free surfaces is small. In this case the interface will behave more like a grain boundary than two separate surfaces. Metal- metal contact is a necessary but not sufficient criterion for galling [12].

(20)

Due to high local pressures in the contact area, the recrystallisation temperature decreased. Frictional heating on micro scale activates recrystallisation of small areas of the interface. In the recrystallisation process material pick up occurs. The frictional heating on a micro scale can be significant, tests has shown that “flash temperatures” on a top of a sliding and deforming asperity to reach several hundred degrees Celsius [12].

Galling initiation and growth in deep drawing

Galling is initiated at imperfections in the tool surface, machining marks and protruding carbides. It is believed to begin at asperity level in the interface between the materials [5]. Most deep drawing operations use lubricants with added molecules that form transfer films. First the transfer film collapses under high pressures and temperatures. The film collapse is facilitated by the plastic deformation of the sheet as well [2]. With the damaged transfer film the asperities now connect in a way that is close to the conditions in dry sliding [5]. The plastic deformation not only damages the transfer film, but it also seriously compromises the integrity of the oxide film, causing cracks and poor adhesion between the oxide and the matrix, with possible oxide film removal. With the oxide removed, the sheet can adhere strongly to the tool material. High pressures and flash temperatures facilitate the forming of adhesive bonds. These bonds become as strong as grain boundaries and the relative motion of the surfaces shears material from the sheet [15].

(21)

In deep drawing galling is the single largest cause of total failure and unacceptable sheet surface roughness [15]. As mentioned before it is critical that the friction is controlled in deep drawing and galling, causing uneven friction, has to be avoided. Galling also creates unacceptable scratches in the sheet.

Heavy duty sheet metal forming operations usually have mixed lubrication because the pressures are too high to promote hydrodynamic lubrication. In mixed lubrication, some of the surface asperities are in contact and slide against each other. The transfer film

protecting the surface is torn as the sheet is plastically deformed, that leads to metal-metal contact. The tendency of transfer film breakdown due to plastic deformation of the sheet is what makes most lubricants work poorly. There are some lubricants based on chloride, sulphur or zinc that are less sensitive to transfer film breakdown and therefore work better in deep drawing.

The contact important in SMF

The most prominent difference in contact situation in SMF compared to other wear contacts is that the tool experience new, virgin sheet surface during its whole life. In other systems, more commonly the same surfaces meet over and over again, allowing for running in.

Surface finish of the tool and roughness of the sheet may influence the products

(22)

Fig. 10. Picture of the surface structure of the sheet and tool, [15].

closer study of the situation reveals that the sheet has rather large islands that protrude on the surface, accounting for the large roughness. The ground tool on the other hand has a much more uniform surface and as the islands are flat, the tool asperities actually

experience a flat counter surface, see Fig. 10. [15]. The tool is often hardened but the sheet is kept soft, resulting in a large difference in hardness. The large difference is to reduce the wear on the tool.

At the die radius the most extreme conditions occur. Here the mean contact pressure can reach 100 MPa and sliding velocities are in the order of 0.5 m/s [15]. The sheet is also deforming plastically reducing the integrity of the protecting oxide layer.

Galling begins with the formation of small lumps on the tool surface. These often form on the side of an asperity or machine grove ridge and grow in the valley. This almost always occurs and is hard to influence [12].

(23)

Fig. 11. Diagram illustrating influence of attack angel on abrasive wear regimes. a) Picture of the different regimes, cutting, plowing and wedge formation. Attack angel of asperities b) under critical attack angel. c) critical attack angel. V is the relative speed of the asperity and the sheet. Θc

is the critical attack angel and Θcw is the critical wedge angel, [15].

Material pickup occurs on asperities in the wedge formation regime, see Fig. 11a. In the

figure fhk is a measurement of the severity of contact and is defined as:

κ τ

= HK

f . Where τ

is the interface shear strength and κ is the shear strength of the softer material. Material is forced up on the asperity in the wedge formation regime exposing virgin metal. In cutting the material is removed from the surface and could then be transferred. This is however unlikely as the removed material quickly forms oxide film, which inhibits adhesion. They also tend to be shovelled down in the valleys. In ploughing, no material leaves the

ploughed surface, thus no pick up takes place [15].

(24)

Methods for galling initiation tests

Several methods to determine materials resistance to galling are developed today. The desire to have a cheep, quick and accurate test has led to several arrangements of test equipment. Here a few of the more commonly used methods are given a short presentation.

ASTM G98.91 standard

Fig. 12. Principal image of the button-on-block configuration, [].

(25)

criterion, it is fast, there is a simple and clearly defined surfaces and it is easy to manufacture samples.

Modified ASTM G98 test

Fig. 13. Principal image of geometry (a), and cross section (b), of the modified G98 test geometry, [17].

The G98 tests almost always show signs of galling at the periphery of the button. This might be accredited to the geometry of the test specimens. The stress in the contact area is assumed to be uniform in the G98 test. Although as mathematical modelling and finite element models showed, there is a significant stress concentration along the edge of the button. In an attempt to reduce the stress concentration influence on the test results a change in the geometry of the button-on-block set up is suggested. The new geometry is that of a hollow cylinder on top of another hollow cylinder, see Fig. 13. The geometry of the alignment pin ensures that the outer and inner edges of both specimens are concentric. The rounded head of the alignment pin also ensures line contact through out the test. Preliminary testing has shown that the tendency of galling to occur at the edge has

(26)

same as with the ASTM G98 test. The modification is only an attempt to increase reliability and accuracy.

The Uppsala Load-Scanner

Fig. 14. Uppsala load-scanner a) principal function of the load scanner, b) test setup, c) typical friction curves from the Uppsala load-scanner. Vanadis and Weartec are tool grade steels and TiN

and DLC are surface coatings, [18].

In the load scanner, two cylindrical specimens, Ø10mm x 100 mm, are positioned in 90˚ angle to each other see Fig. 14 a,b. They slide in such a way that at the end of the stroke both the specimens are at the ends [18]. The load is increased during a forward stroke, typical from 30 to 1300N. The friction force and normal force are logged during the test. After the test the coefficient of friction is examined and a sudden increase of the

(27)

contact longer then the size of the contact area [18]. The disadvantage is an extremely short sliding distance. The advantages are easy sample manufacturing, few samples needed as the same samples can be used more than once and a history of the run is represented in the specimens.

TNO Tribo Tester

Fig. 15 a) Principal sketch of the TNO Tribo Tester, b) typical friction curve from the TNO Tribo Tester. This test was performed with lubrication, [19].

(28)

the test is run on a limited amount of sheet metal with a simple geometry and a long sliding distance. Real sheet is used.

Bending Under Tension, BUT

Fig. 16 a) Principal sketch of the BUT rig, b) close-up of the bending radius, c) typical results from a BUT test, [21].

In this test an effort has been made to simulate the real situation. The BUT test uses a strip of sheet metal that is put under strain and then slid over a rounded 90˚ corner, see Fig. 16a. This is then repeated many times with fresh sheet each time. The sliding distance of

(29)

The Olofström U-bending test

Fig. 17 a) principal sketch of the U-bending rig, b) close-up of the draw radii inserts, c) typical acoustic emissions from a U-bending test, [22].

(30)

Uddeholm Tooling Tribo Tester

The UTTT is a cylinder on cylinder test with sheet metal. In this test a small cylinder, Ø 10 mm, of tool material is slid against a sheet. The sheet is placed on a larger cylinder, Ø 300 mm, and is pre-tensioned. A constant load is applied on the tool cylinder and the sheet is rotated so that a sliding distance of 750 mm is achieved. The test is repeated with a higher load until galling occurs. Then the average of the highest non galled load and the lowest galled load is calculated. This is the “galling load” at which the pared materials begin to show signs of galling. During the test forces on the test cylinder are recorded and a sudden increase in the coefficient of friction is the galling criterion. The sheet is also examined for galling. The disadvantages are that the sheet tension is not measured, the sliding distance is short and a lathe is required. The advantages are easy sample manufacturing few samples needed as the same samples can be used more than once and it is relatively cheep and fast (provided you have a lathe). Real sheet is used.

Aims and scope

Seeing that galling is a problem in sheet metal forming the need for an in house galling test for Uddeholm Tooling AB is clear. In order to provide customers with accurate and

relevant information a ranking of tool steel resistance to galling is imperative. Support as to how to get the most out of the products depends on the knowledge of the company’s research.

The current galling test setup at Uddeholm Tooling AB has an obvious upgrade potential and a few improvements are addressed in this work.

(31)

The second update was to design and implement a new tool holder. The tool holder used is rough and is prone to jam during constant load tests. The new holder has to be able to handle these tests as the steel sheet holds a much shorter continuous sliding distance. The shorter sliding distance makes increasing load tests harder to perform. To modify the galling tester to allow for multiple rotations is not included here but is a development encouraged as future work.

The third update was to design and implement a new data acquisition system (DAQ-system) Parts of the old system is kept but the acquisition board, computer and software is changed. The main reason for this is to make the data easier to handle and to be able to apply basic filtering to the data.

When the improvements have been put into operation the galling test rig is evaluated. The evaluation is done to determine if a ranking of tool steels resistance to galling is obtainable by the improved setup.

Materials and Methods

In sheet metal forming there are basically two different material groups, the sheet material and the tool material. The sheet is soft and ductile and has a rough surface whereas the tool is hard and smooth. Chemical analysis and specimen specifications are given in Table 1 and Table 2 respectively.

(32)

Table 1. Nominal chemical composition of sheets and tool steels in wt. % Steel grade C Si Mn P S Cr Ni Mo V Ti DC 04 0.016 0.001 0.18 0.009 0.008 0.03 0.04 <0.01 0.002 0.0005 DOCOL 1000 0.16 0.45 1.45 0.013 0.004 0.04 0.04 <0.01 0.01 0.0023 AISI 304-10 0.036 0.44 1.21 0.028 0.001 18.5 8.79 0.25 0.05 0.007 UNIMAX 0.49 0.08 0.5 0.01 0.0007 5.0 0.2 2.4 0.56 -SVERKER 21 1.55 0.17 0.3 0.02 0.01 11.1 0.1 0.7 0.77 -Steel grade Cu Al Ce B Ca Co W N Nb DC 04 0.007 0.028 0.002 0.0001 <0.0006 0.015 0.01 0.014 -DOCOL 1000 0.006 0.039 0.002 0.0001 0.003 0.015 0.003 0.010 -AISI 304-10 0.40 0.003 0.011 0.0001 0.0012 0.15 0.02 0.064 -UNIMAX 0.1 - - - - 0.05 0.07 - 0.003 SVERKER 21 0.09 - - - - 0.04 0.07 - 0.006

Table 2. Tool steel test specimen specifications

Steel grade Sverker 21 Unimax

Bulk material dimension [mm] 250 x 80 250 x 80

Cut to lengt [mm] 75 75

Specimen orientation LT LT

Finished dimensions [mm] Ø 10 x 60 Ø 10 x 60

Ra < 0.1 < 0.1

Hardness HRC 60.7 57.5

BalinitC Surface coating on Sverker 21 specimens.

The sheet materials used are AISI 304-10, DC04 and Docol 1000 DP. AISI 304-10 is an austenitic stainless steel. These steels are notorious for their poor galling resistance. DC04 is mild steel used in pressing and deep drawing operations. Docol 1000 DP is a high strength dual phase steel with good formability.

Sample preparation

(33)

they were hard turned to the correct shape and hand ground with 1200 paper until Ra < 0.1 μm. The samples were cleaned first in acetone and then alcohol and left for at least 3 days to allow a stable oxide film to form. The sheet is cut to length from a roll and 3 holes are drilled in each end. It is then strapped down on to the sheet holder and tensioned with screws. The sheet is cleaned with acetone and alcohol and left for at least one day to give it time to settle on the sheet holder. Immediately before each test the specimen and the sheet are cleaned with acetone and a soft cloth to remove any oil based contaminants. Then the cleaning proceeds with alcohol to remove residues of acetone and any hydro based contaminants.

The UTTT rig

A model test

The choice of test method must be done carefully. A correctly chosen test saves money and time and gives accurate results. The fist thing to consider is what class the test should be. The most realistic class is the field test in which the entire system is run in its natural environment. In SMF this would correspond with a production line type of machine actually producing products in the factory. The next best thing is the bench test were a production line type of machine in a laboratory produce products. In the laboratory the surrounding environment can be controlled and manipulated. In the part system class the system is reduced to the most significant parts, for instance the punch, die and guides. Often vibrations and environment differ significantly from the field test. In the

component test class the parts of interest is taken from the production line machine. The

die and punch are used to produce products under controlled conditions. In a simplified

component test class only part of the components of interest is tested. For instance only

(34)

components have been replaced. Instead of a punch and die only the tool material is used against the sheet material. The sheet can be replaced with a solid geometry of the same material.

The field test produces the most reliable results but is time consuming, expensive, hard to control and varies with the operator. The model test, on the other hand, has a high degree of control, reproducible results. This test is fast and cheep but is often far removed from the real application with respect to geometry, environment, time and contact pressure, [1]. UTAB need to be able to recommend tool steels to their customers and different

customers have different applications. The model test class is the best suited, this test class will give a general guide to the pairs of materials that work best.

The model test class is large and includes many tests. The large numbers of tests become more manageable if they are ordered into sub classes. One way of doing this is to differ between open and closed tests. In an open test at least one of the surfaces meets fresh counter material during the entire test. In a closed test the surfaces meet over and over again. The cylinder on cylinder test is a test that can be either open or closed. The cylinder on cylinder test consists of two cylinders loaded perpendicular to each other, one cylinder is then rotated. The cylinder being rotated is here referred to as “the roll” and the non-rotating simply “the specimen”.

In the closed test mode the roll is rotated numerous laps and the specimen is kept

stationary. The same spot on the roll will come in contact with the specimen each rotation. The wear track on the roll would be that of a single ring, no matter how many rotations are made.

In the open test mode is the specimen on the other hand moved along the length axis of the roll as it rotates. During the run the specimen will always meet a fresh surface. The wear track would coil its way up the roll like the threads on a bolt.

(35)

whether the specimen is held stationary or not. Because a full rotation is never completed the specimen will constantly meet a fresh counter surface.

Other ways of divide model tests are with respect to wear dependence. If during the run the contact area is allowed to change due to wear as in the cylinder on cylinder test the test is a wear dependent test. If the contact area does not change as it is worn like when a flat pin is pressed against a flat rotating disc the test is wear independent, [1].

A cylinder on cylinder, wear dependant, open mode, model test has been selected. In order to get a general result that would be applicable to many products a model test seems the best choice. It also has the advantage that tests can be made relatively fast and at low costs. The tools used in deep drawing often have a wear dependent geometry and thus so should the test. Seeing that the tools in production are in contact with virgin sheet with every stroke and no or little running in is allowed an open mode test seems appropriate.

To chose a criterion for galling

There are a few ways to determine that galling has occurred. The most popular way is to monitor the friction under the test. Typically, the materials have a short running in period where the friction is unstable. This period is then followed by a steady run, the working phase, where the friction is low and stable. Next to the end of the tool life a sudden

increase and unstable behaviour of the friction is observed. This sudden increase is used as a galling initiation criterion [23]. The method is widely used and has the advantage that it can be read continuously during the test.

Instead of looking directly on the friction curve a study of the standard deviation of that curve can be used to judge how unruly the signal is. The standard deviation of the

(36)

Another way is to look on transferred material on the tool surface [15]. In this case the number of lumps, height or the area they cover is noted. This can be done by measuring or more subjectively by visual inspection of the surface and grading against a reference scale. The disadvantage of this approach is that the measurement can not be done continuously during the run.

A third method is by inspecting the sheet. Galling causes scratches in the sheet and the galling criterion is when these scratches become unacceptable. The scratches are either measured or visually inspected, [13,22].

Fig. 18. The old UTTT setup, a) the solid roll with the machining arm, b) the old tool holder, c) diagram showing the fluctuation of the old sheet holder at a load of 100 N. Fz is the concern.

(37)

Hardware

The existing test rig consisted of a solid roll, see Fig. 18a. 150-160 mm Ø of the sheet material mounted in a lathe (Köpings Mekaniska Verkstad Typ S10C Nr 596). The tool material was mounted in a simple holder, see Fig. 18b. This set up functioned poorly.

The sheet material

Fig. 19. The new sheet holder.

(38)

holder. In an area round the screws the sheet is not following the holder accurately and this part of the sheet must be avoided. The usable length of the sheet is 750 mm. In using real sheet material the surface has the typical topography, the hardness is the same along the entire sheet and the surface has a stable oxide layer. The lathe rotates at a speed of 14 rpm giving the sheet a velocity of 0.22 m/s. The new sheet holder does have its limitations, as the sheet is discontinuous runs are limited to a single rotation. The strain in the sheet is not measured and is bound to vary from sheet to sheet.

The tool holder platform

The simple tool holder consisted of two cylinders, one inside the other with a neoprene bushing between. Inside the larger cylinder one or two springs were placed and the smaller cylinder compressed the springs when a load was applied. This tool holder is prone to wedging during constant load tests. The offset of the solid roll and the spring constant has been measured to approximately 0.04 mm and 110 N respectively. This yields a theoretical fluctuation of the normal load of 4.4 N/rotation at 100N. Measurements of the normal load have shown a fluctuation of 25 N/rotation at 100N, see Fig. 18c. Another concern was that the holder left markings the tool specimen.

(39)

Fig. 20. The new tool holder platform, a) CAD assembly, b) photograph, c) different tool holders, concept CAD assembly.

The Data Acquisitioning System

The data acquisitioning system or DAQ-system consists of a three-component measuring platform, a 3 channel amplifier, a BNC connection board, a DAQ-board, a PC and software.

(40)

5000 N plus 50% overload. More information on the Kistler platform can be found in Table 4. The amplifier is a Kistler Charge Amplifier typ 5001 with XYZ-compensator type 5201. The BNC connector board is National Instruments BNC-2120 and the DAQ-board is National Instruments PCI-6221.

Software

The software has been developed in National Instruments LabVIEW 8.2 in order to collect and process the signals. The software has to be able to display the signals in close to real time, perform a zero calibration, accept user settings, perform the measurement, display the coefficient of friction and its standard deviation, filter the signal according to user

(41)

Fig. 21. The Software, front panel.

filtering is performed. In the seventh tab the user can compare the current signal to a previously saved one.

To be sure that the displayed signal is the correct signal some basic filtering is done on the signal. To avoid over tones the signal is sampled with more than double the desired resolution. It is then passed trough a low pass filter with an upper limit at the desired resolution. The over sampling ensures that for every frequency up to the desired resolution at least two samples per oscillation are taken. Two samples per oscillation are enough to be certain that the sampled signal is correct. The filtering then removes all the frequencies above the desired resolution and along with them possible overtones. This filtering is hidden inside the program and can not be altered by the user.

(42)

Results

The UTTT have been improved and tests run on real sheet mounted in the new sheet holder. The sheet holder allows for runs of nearly a full rotation with continuous contact of 750 mm. It does not however accommodate multiple rotations at its present state.

Fz -10 0 10 20 30 40 50 60 70 80 90 100 110 120 9,5 10 10,5 11 11,5 12 12,5 13 13,5 14 14,5 S Ne wton Z

Fig. 22. The new tool holder is capable of holding a stable force that varies only within the 5N that is to be expected from the round-out of the sheet.

The new tool holder platform is in operation and has the required stable Fz curve, see Fig. 22. The platform has also proved easy to operate and is flexible. The new DAQ-system and software has been implemented and data can now be filtered at site and exported as a Microsoft Excel file. A problem with real time display of the signal remains to be attended to, as towards the end of measurements the plot lags behind.

(43)

Fig. 23 Micrograph of the sheet surface topography after testing, a) AISI 304-10 Sheet- From the bottom the applied load is 20, 30, 40 and 50 N. b) DC04 Sheet. From the bottom the applied load is

40, 50, 50 and 20 N. c) DOCOL 100 DP Sheet. From the bottom the applied loads are 50, 60, 60 and 50 N. The Tracks made at 50N are hardly distinguishable whereas the tracks made at 60N are

clearly visible.

done with SVERKER21 and loads down to 20 N, at which the specimens showed adhered material under a visual inspection.

(44)

Fig. 24 Micrograph of the tool surface topography after testing, a) BalinitC specimen. The dark area is the wear scar and there is clearly no or very little adhered material present. This test was done with a load of 50N and is the mate of track (i) Fig. 23c. b) BalinitC specimen. The dark area is the wear scar and there clearly is adhered material in the centre. The transferred material is also clearly visible to the naked eye. This test was done with a load of 60N and is the mate track (ii) Fig.

(45)

The coefficient of friction begins to increase from 0.2 to 0.4 around 50 N for BalinitC see Fig. 26a. The increase is rather rapid and occurs in the span of 50-100 N. SVERKER21 and UNIMAX have a high coefficient of friction even at the lowest load and has only a slight tendency to increase at higher loads.

The standard deviation of the coefficient of friction increased drastically at low loads, see Fig. 26b. BalinitC seems to be more stable at the lower loads than SVERKER21 and UNIMAX even though it to increases.

Scratches in DOCOL 1000 DP

0 1

0 50 100 150 200 250 300 350 400 450

Applied load in Newton

Scrat c h i n t h e sh eet : Y es = 1 / N o = 0 SVERKER21 BalinitC UNIMAX

(46)

a)

Coeficion of friction vs Applied load on DOCOL 1000 DP 0,000 0,100 0,200 0,300 0,400 0,500 0,600 0,700 0,800 0 50 100 150 200 250 300 350 400 450

Co ef icien t o f fr ict ion Sverker21 BalinitC Unimax b)

Standard deviation of the coefficient of friction vs Applied load. DOCOL 1000 DP -0,100 -0,050 0,000 0,050 0,100 0,150 0,200 0,250 0,300 0,350 0,400 0,450 0 50 100 150 200 250 300 350 400 450

ST

D Sverker21

BalinitC Unimax

Fig. 26. Data from tests on the UTTT, a) diagram showing a decrease in the coefficient of friction for BalinitC, b) diagram showing the standard deviation of the coefficient of friction plotted against

(47)

Discussion

Galling is a type of surface damage containing a large extent of adhesive wear.

Nevertheless, classification of adhesion-related wear mechanisms is not very clear by now. There are, mild, severe and catastrophic adhesive wear, Type I, II, III, IV and V sliding-rolling wear, seizure, scuffing, scoring, galling and gouging.

The terms seizure, scoring, scuffing and galling has previously often been associated with the physical environments. Seizure for example is regularly linked with pure, clean surfaces worn in vacuum. Catastrophic adhesive wear under lubricated conditions has been referred to as scuffing or scoring. Galling has been used to express adhesive wear under unlubricated conditions. In [24] it has been proposed to give a more physical mechanism-specific explanation to the phenomena behind the names. Common for these adhesive wear types are plastic flow or ploughing of at least one of the surfaces and concentration of normal load to a small area in the contact. Common is also the appearance and growth of protrusions and groves resulting in a roughening of the surfaces and accelerated wear. In seizure only one surface is deformed plastically and material is adhesively transferred from the softer surface to the harder on a macroscopic scale. The hard surface develops hard protrusions that cause groves in the softer material. Scoring does not have adhesively transferred material on a macroscopic scale but both surfaces are deformed plastically. The plastic deformation causes the material to work harden and form wedges on either surface. In scuffing material is welded to one of the surfaces as a consequence of frictional heating and softening of one surface. Galling first transfers soft material that has not yet been heated and smears it on the hard counter surface. Once on the counter surface the smeared material undergoes significant heating.

(48)

Table 3.

high-energy types because of the higher sliding speeds required to achieve the frictional heating necessary. The sliding speed necessary to obtain the high-energy types is estimated to above 0.7 m/s for carbon steels by [24], see Table 3. Note that [24] definition of galling does not mention the presence of macroscopic lumps on the harder surface. In fact it would seem that the term seizure corresponds better to the phenomena studied here. According to this table high energy adhesive wear occur only at speeds above 0.4 m/s. In this thesis a speed of 0.22 m/s was used and therefore is results directed to the low energy types. In the present research the morphology of worn surfaces are similar to those described in the literature. The tool specimens clearly show adhered material and this rule out scoring, which has no macroscopic transferred material. In Fig. 24. it is clearly the beginning of formation of lumps on the tool and in Fig. 23 groves are clearly visible in the sheet. Fig. 24 shows small lumps and adhered material on the tool surface. It could be argued that this is not what [11] has defined as galling, if the lumps and adhered material on the tool do not grow any further. However at increased loads every one of the tool specimens showed large amounts of adhered material, see Fig. 27, consistent with [11] definition.

(49)

Fig. 27. Typical appearance of adhered material on SVERKER21 at loads higher than 100N, Pictures taken by V.B. SARMA

would complement but probably not render any new conclusions.

On the DC04 sheet vs. SVERKER21 (20 – 700 N) show a dip in the coefficient of friction at the low loads. This could indicate that the transition load is close to the lower loads but as all tests have shown scratches and transferred material it is not certain. Tests with BalinitC were not performed on DC04. The dip in coefficient of friction for

SVERKER21 could indicate that the sheet is less sensitive to galling compared to DOCOL 1000 DP but without tests that can indicate an upper limit no conclusions can be made. The DOCOL 1000 DP tests show a transition load for BalinitC at 50-60 N. There is however few tests done at higher loads and more tests could help to decrease the error. SVERKER21 has a small dip in coefficient of friction at low loads but not near enough to be called a transition. This is also confirmed in that scratches and adhered material was always present. The few tests done with UNIMAX is explained by the fact that even the lowest load test scratched the sheet and resulted in material pickup. Testing was aborted quickly.

Investigating scratches on the sheet is important to distinguish between a track formed by plastic deformation and a scratch formed by galling. The former is normal and is

(50)

magnifier. The tool specimen was examined under microscope and for all runs that have resulted in a scratch material pick up was noted on the tool. The runs that resulted in a track corresponded with no transferred material on the tool specimen. This might not be sufficient to establish galling but it is a strong indication of adhesive wear. Other ways of dealing with this problem is to measure the scratches with a profilometer and determine a maximum value that is acceptable to the manufacturer. In doing this the plastic

deformation of the sheet should be mathematically removed otherwise a larger plastic deformation will cause an increase in the profile depth.

(51)

the point for galling initiation tends to differ between runs. The interpretation of the data is more sensitive with the Uppsala Load-Scanner because only a small amount of the data can be liked to a specific load.

In the initial phase, before galling, the coefficient of friction was observed to be between 0.3 and 0.4 in the Uppsala Load-Scanner, see Fig. 14c. The UTTT shows a coefficient of friction of between 0.4 and 0.6 for galled specimens indicating that if galling had not occurred the coefficient of friction should be below 0.4. The low friction coating of BalinitC showed an initial coefficient of friction of around 0.2 that is comparable to the low friction coating DLC’s coefficient of friction of 0.15 in the Uppsala Load-Scanner. The TNO Tribo Tester uses lubrication and achieves a lower coefficient of friction even though it behaves much the same way in that it increases and eventually becomes erratic and unstable, see Fig. 15b. The UTTT also displays an increase during the transition from no galling to galling. However, the coefficient of friction in the UTTT tests behaves contrary to the expected with regards to the standard deviation. The coefficient of friction displays erratic behaviour at low loads and becomes more stable during higher loads. The expected behaviour is a steady coefficient of friction before galling occurs and erratic and unstable coefficient of friction after galling has occurred. The unexpected behaviour of the coefficient of friction is attributed to the Kistler measuring device, since it is designed to measure larger loads.

(52)

profilometer. Measurements of the wear track are taken in the TNO Tribo Tester and the roughness of the sheet has been found to display the same characteristics as that of the coefficient of friction. This is more time consuming but has the advantage that it is objective and the data is more easily compared with previous readings. For more information on measuring wear tracks see [13].

The contact area in the UTTT has the shape of a small ellipsoid and is dependant on the applied load and the type and amount of wear that occur. The load dependant contact area characteristic is shared with the Uppsala Load-Scanner and the TNO Tribo Tester. The advantage of this type of contact is its robustness in regard to specimen tolerances and the lack of sharp edges. The ASTM G98 tests are more sensitive to specimen shape because small differences in specimen geometry could have rather large consequences. The problems with sharp edges are more of a practical nature, because sliding against a sharp edge results in shear rather than galling. Even if the geometry allows for sharp edges they should be treated with care as the modified ASTM G98 test clearly states.

A wear dependant contact changes during a test and the contact area might grow and thereby reducing the contact pressure and temperature. A wear independent contact often has practical problems with geometry and edges. The best contact situation is found in the BUT and Olofström U-bending tests as they closely simulate the contact situation in sheet metal forming.

(53)

outperforms the equipment that has a maximum contact pressure of 5.2 GPa [18]. Perhaps the difference between the test setups is enough to explain this. The TNO Tribo Tester uses lubricated sheet and the Uppsala Load-Scanner uses machined sheet material. The dry sliding condition on real sheet could significantly lower the galling pressure.

The UTTT use real sheet material to better simulate the conditions in SMF. This has the advantage of supplying a sheet material with the surface structure used in fabrication. Because the UTTT tests were done under dry sliding conditions the corrosion protection had to be removed. The oil coating was cleaned with acetone and then alcohol to produce a relatively clean surface. To remove the protective coating completely would require special cleaning baths and ultra sonic treatment. The test methods that use machined sheet metal specimens do not have to remove a protective film since it has never been applied, but there might be other contaminants present. To use real sheet is a major upgrade since the surface quality influence the wear process [18]. To use real sheet metal might have its limitations because it is harder to experiment with parameters such as surface roughness with real sheet. In most instances this is of little or no concern as the surface quality not obtainable in a real sheet is not going to be involved in any SMF.

(54)

sliding distance would be dramatically improved. This would allow for a single test to spend the entire strip of sheet, resulting in sliding distances up to 40 x 750 mm = 30 m. This is an impressive boost in sliding distance without increase in the dimensions of the sheet. The sliding distance would be in mid range, longer then the ASTM G98 tests but shorter then the TNO Tribo Tester, while at the same time require considerably smaller amounts of sheet then the BUT or Olofström U-bending tests.

The unsteady behaviour of the coefficient of friction at low loads might be attributed to the data acquisition hardware. A Kistler 3-axis force measuring device was used to record the friction and normal forces. The Kistler platform has a maximum load capacity of 5000 N plus 50% over load, placing the low loads of 10-100 N used in the tests at the very end of the scale. It is far better to have the load range in the middle of the measurement device’s range.

In this thesis no lubrication were applied to the sheet or the specimens. The only exception being the low friction coating BalinitC that arguably could be termed as

(55)

that lubrication would bring. If the threshold load increased above 100 N the Kistler platform would perform better and give more consistent readings.

Future Work

The most important change to the UTTT would be a way to measure the strain of the sheet. The operator tightens the sheet manually and the strain in the sheet differs from sheet to sheet. The sheet tension is most likely to influence the results of a test and should be kept constant during a series of tests. A known and constant tension would give less scatter in the results and greatly improve repeatability of tests. To monitor the sheet strain a strain gage could simply be glued directly on the sheet before it is placed in the UTTT.

To replace the Kistler platform for a smaller one with a measuring range of 0.1 - 250 N would probably reduce the standard deviation of the coefficient of friction. This may require modifications to be made to the tool holder platform as well. New fittings to secure the tool holder platform to a new measuring device must be made. The tool holder platform would also benefit from a change from ball cages to bushings for a smoother force transfer at low loads. To change the Kistler platform is expensive and it is not certain that it solves the problem, the large standard deviation of the coefficient of friction at low loads could originate from vibrations in the lathe.

The current geometry of the tool steel specimen could be modified in order to allow a higher load to be applied without increasing the contact pressure. The problem with a large standard deviation of the coefficient of friction at low loads would be solved if the loads could be increased. A change in geometry that gives the same contact pressure but at loads in the 200-1000 N range could make the Kistler platform perform more stable.

(56)

results and are easy to compare. To measure the scratches is not as fast as visual inspection. The sheet strip is easily removed after testing and can be cut for closer examination.

One easy modification would be to switch from dry sliding to lubricated runs. The switch can be done without any modifications to the hardware. By simply applying lubrication to the sheet before testing will suffice. More elaborate ways to lubricate the sheet might be devised to better control the lubrication. Lubrication should reduce the tendency of galling and because of this the loads required to achieve galling might be high enough for the Kistler platform.

A modification to the test rig to allow multiple revolutions would also be a worth while. In order to permit multiple rotations the tool steel specimen needs to be lifted from the sheet in so that the seam in the sheet may pass untouched under. This can easily be done with a cam and cam follower that lifts the tool holder from the sheet. This modification requires extra pieces of hardware to be manufactured and fitted. The increase in sliding distance possible by this would be substantial, from 0.75 m to 30 m. This longer sliding distance should go well together with lubrication.

The UTTT is a test with a good foundation and a great upgrade potential. The test uses real sheet and thereby simulates the situation in SMF well. The sliding speed is in the range used for SMF and the distance of continuous sliding is good. The tool steel

(57)

Conclusions

Only for BalinitC has a galling load been detected. The other materials have all galled even at the lowest load possible with this test setup. Even BalinitC has a very low galling load. From this the only conclusion to be drawn at this stage is that the span is offset down in the measuring range. This makes the current test setup unable to rank tool steles to their

sensitivity towards galling.

Acknowledgements

Following persons deserve special thanks. Åslund, Linda for unwavering support and love.

Krakhmalev, Pavel, supervisor at KAU, for patience, help and guidance.

Nordh, Lars-Göran supervisor at Uddeholm, for plenty of discussions and ideas. Gunnarsson, Staffan supervisor at Uddeholm, for keeping an eye on the project. Lars for operating the milling machine and for practical knowledge in design. Lennart for operating the turning machine and for practical knowledge in design. Högman, Berne for discussions and help with software and programming.

(58)

Appendix A

Table 4 The Kistler platform data sheet

type

Nr.

Ed.

P. Kistler 9257A B6.9257e 5.71 1…18

Max measuring range 5000 N

Max measuring range when turning 10000 N

Resolution <0.01 N

Overload capacity 50 %

Sensitivity Fz -3.5 pC/N

Fx, Fy -7.5 pC/N

Stiffness in Z axis 350 N/μm

Stiffness in X and Y axes 1000 N/μm

Resonant frequency abt. 4.0 KHz

Linearity (max error) <±1 %

Crosstalk < 2 %

Insulation resistance (each channel) > 10E13 Ω

Capacitance (each channel) 140 pF

Temperature coefficient -0.02 %/?C

Operating temperature -100/+100 ?C

Weight abt. 5 kg

Aberrations: SMF Sheet Metal Forming BUT Bending under tension

ASTM American Society for Testing and Materials

PLP Preservation, lubrication, and Primer-Compatibility

DAQ Data Acquisition

HRC Hardness Rockwell C-scale

(59)

Table 5. Test matrix.

Tool Sheet [N] Tool Sheet [N]

BalinitC DOCOL 1000 DP 20 SVERKER21 DC04 20

BalinitC DOCOL 1000 DP 350 SVERKER21 DOCOL 1000 DP 10

BalinitC DOCOL 1000 DP 400 SVERKER21 DOCOL 1000 DP 10

SVERKER21 AISI304-10 20 SVERKER21 DOCOL 1000 DP 15

SVERKER21 AISI304-10 300 UNIMAX DOCOL 1000 DP 10

(60)

References

[1] Jacobson, Staffan. Tribologi: friction, smörjning och nötning. Stockholm : Liber utbildning 1996

[2] Haar, Rudi ter. Friction in sheet metal forming, the influence of (local) contact conditions and deformation. Enschede : University of Twente, 1996

[3] Callister, William D. Jr. Materials science and engineering an introduction. New York : Wiley, 2003

[4] http://www.zygo.com/library/papers/SurfText.pdf, 2007-04-18

[5] Heide, Emile van der. Lubricant failure in sheet metal forming processes Enschede : University of Twente, 2002

[6] Jayadas, N.H. Tribological evaluation of coconut oil as an environment-friendly lubricant, in Tribology International. 2007, 40, pp.350-354

[7] On the friction and wear performance of boric acid lubricant combinations in extended duration operations, in Wear. 2006, 260, pp.1295-1304

[8] Lovell, M. Mobley Increasing formability in sheet metal stamping operations using environmentally friendly lubricants, in Journal of Materials Processing Technology. 2006, 177, pp.87-90

[9] Roescher, Arno. A new coating for deep drawing with preservation-lubricant-primer properties, in Progress in Organic Coatings. 2001, 43, pp.111-122

[10] Westeneng, André. Modelling of contact and friction in deep drawing processes, Enschede : FEBO druk B.V., 2001

[11] Standard Terminology Rating to Wear and Erosion, G40, Annual book of ASTM standards, vol 03.02, ASTM

[12] Totten, George E. Mechanical Tribology. New York, Marcel Dekker Inc, 2004 [13] Andreasen, J.L. Quantification of Galling in Sheet Metal Forming by Surface

Topography Characterisation, in Int. J. Mach. Tools Manufact. 1998, Vol 38, Nos 5-6, pp.503-510

[14] Heide, E. van der. Galling initiation due to frictional heating, in Wear. 2003, 254, pp.1127-1133

(61)

[16] Hummel, S.R. An application of frictional criteria for determining galling thresholds

in line contact tests, in Tribology International. 2002, 35, pp.801-807

[17] Waite, R.A. Analysis of the stress field in a threshold-galling test, in Tribology International. 2006, 39, pp.1421-1427

[18] Podgornik, B. Comparison between different test methods for evaluation of galling properties of surface engineered tool surfaces, in Wear. 2004, 257, pp.843-851 [19] Heide, E. van der. The effect of lubricant selection on galling in a model wear test, in

Wear. 2001, 251, pp.973-979

[20] Carlsson, Per. Tribological behaviour of thin organic permanent coatings deposited on hot-dip coated steel sheet – a laboratory study, in Surface and Coatings Technology. 2000, 123, pp.169-180

[21] Alinger, M.J. Evolution of the tribological characteristics of several forming die materials, in Journal of Materials Processing Technology. 2001, 111, pp.20-24 [22] Skåre, T. Wear and frictional behaviour of high strength steel in stamping monitored

by acoustic emission technique, in Wear. 2003, 255, pp.1471-1479

[23] An atomistic approach to the initiation mechanism of galling, in Computational Materials Science. 2006, 37, pp.193-197

[24] Markov, D. Mechanisms of adhesion-initiated catastrophic wear: pure sliding, in Wear. 2000, 239, pp.189-210